Novel Respiratory Syncytial Virus-Like Particle Vaccine Composed ofthe Postfusion and Prefusion Conformations of the F Glycoprotein

Velasco Cimica,a Hélène Boigard,a Bipin Bhatia,b John T. Fallon,c Alexandra Alimova,d Paul Gottlieb,d Jose M. Galarzaa

TechnoVax, Inc., Tarrytown, New York, USAa; EMD-Millipore Corporation, Billerica, Massachusetts, USAb; Department of Pathology, New York Medical College,Westchester Medical Center, Valhalla, New York, USAc; Department of Pathobiology, Sophie Davis School of Biomedical Education, City College of New York, NewYork, New York, USAd

Respiratory syncytial virus (RSV) is the leading cause of severe respiratory disease in infants and children and represents an im-portant global health burden for the elderly and the immunocompromised. Despite decades of research efforts, no licensed vac-cine for RSV is available. We have developed virus-like particle (VLP)-based RSV vaccines assembled with the human metapneu-movirus (hMPV) matrix protein (M) as the structural scaffold and the RSV fusion glycoprotein (F) in either the postfusion orprefusion conformation as its prime surface immunogen. Vaccines were composed of postfusion F, prefusion F, or a combina-tion of the two conformations and formulated with a squalene-based oil emulsion as adjuvant. Immunization with these VLPvaccines afforded full protection against RSV infection and prevented detectable viral replication in the mouse lung after chal-lenge. Analyses of lung cytokines and chemokines showed that VLP vaccination mostly induced the production of gamma inter-feron (IFN-�), a marker of the Th1-mediated immune response, which is predominantly required for viral protection. Con-versely, immunization with a formalin-inactivated RSV (FI-RSV) vaccine induced high levels of inflammatory chemokines andcytokines of the Th2- and Th17-mediated types of immune responses, as well as severe lung inflammation and histopathology.The VLP vaccines showed restricted production of these immune mediators and did not induce severe bronchiolitis or perivas-cular infiltration as seen with the FI-RSV vaccine. Remarkably, analysis of the serum from immunized mice showed that the VLPvaccine formulated using a combination of postfusion and prefusion F elicited the highest level of neutralizing antibody andenhanced the Th1-mediated immune response.

Human respiratory syncytial virus (RSV) is the leading cause ofsevere pediatric pulmonary disease worldwide. RSV infects

nearly all infants at least once by the age of 2 years. Epidemiolog-ical studies around the globe indicate that 2 to 5% of the childreninfected with RSV require hospitalization, with the most severemorbidity and mortality disproportionally seen in premature in-fants. RSV disease causes 100,000 to 200,000 fatalities per yearglobally (1, 2). It is believed that severe RSV infection can predis-pose children to develop wheezing with future illnesses and po-tentially to develop asthma (3, 4). RSV infection elicits neutral-izing antibodies and a T-cell response that wane over time;consequently, the patient is often unprotected against reinfec-tion (5, 6). Furthermore, elderly people show a greater risk ofsevere RSV disease upon reinfection (7). Despite decades ofresearch efforts, no licensed vaccine is currently available tocontrol or prevent RSV infection (8). Vaccinology researchshows that the F glycoprotein is the most attractive target foreliciting neutralizing antibodies against the virus. RSV displaysdifferent conformations of F that are antigenically distinct: thehighly stable postfusion and the metastable prefusion (9).Magro et al. (10) have demonstrated that antibodies specific toprefusion F account for most of the neutralizing activity in aprophylactic human Ig preparation and immunized rabbits.Subsequently, McLellan and coworkers (9) determined theprotein structure of the prefusion F by X-ray crystallographyand identified the prefusion-only antigenic site � (Fig. 1A).While palivizumab can recognize both postfusion and prefu-sion structures, a subset of highly neutralizing antibodies (5C4,AM22, and D25) bind specifically to the prefusion antigenicsite � (9, 10). Interestingly, the AM14 and MPE8 neutralizingantibodies are also able to very efficiently recognize the prefu-

sion F using alternative antigenic sites. This demonstrates thatthe prefusion F expresses multiple epitopes suitable for targettherapy (11, 12), which are not exhibited in the postfusionconformation.

Adopting structural vaccinology, our group has developed vi-rus-like particle (VLP) vaccines containing recombinant postfu-sion and prefusion F hybrids together with the human metapneu-movirus (hMPV) matrix protein (M). Efficacy studies showedthat immunization with prefusion F VLP vaccine, postfusion FVLP vaccine, or a combination of the two VLP vaccines (comboVLP vaccine) afforded complete protection against an RSV chal-lenge. Importantly, VLP vaccination was safe and effective in stim-ulating a Th1-type cytokine profile and the combo VLP vaccineelicited the highest level of IgG2a antibody and neutralization ac-tivity.

Received 23 December 2015 Returned for modification 24 January 2016Accepted 19 March 2016

Accepted manuscript posted online 30 March 2016

Citation Cimica V, Boigard H, Bhatia B, Fallon JT, Alimova A, Gottlieb P, Galarza JM.2016. Novel respiratory syncytial virus-like particle vaccine composed of thepostfusion and prefusion conformations of the F glycoprotein. Clin VaccineImmunol 23:451– 459. doi:10.1128/CVI.00720-15.

Editor: M. F. Pasetti, University of Maryland School of Medicine

Address correspondence to Jose M. Galarza, jmgalarza@technovax.com.

Supplemental material for this article may be found at http://dx.doi.org/10.1128/CVI.00720-15.

Copyright © 2016 Cimica et al. This is an open-access article distributed under theterms of the Creative Commons Attribution 4.0 International license.

crossmark

June 2016 Volume 23 Number 6 cvi.asm.org 451Clinical and Vaccine Immunology

on June 23, 2020 by guesthttp://cvi.asm

.org/D

ownloaded from

MATERIALS AND METHODSStructural vaccinology. The RSV fusion (F) (GenBank accession numberACO83302.1) and human metapneumovirus matrix (M) (GenBank ac-cession no. AIY25728.1) genes were codon optimized and chemically syn-thesized (Blue Heron Biotech, WA). Prefusion F mutants were designedby protein structure analysis using Cn3D software (NCBI, MD) and datafrom the NCBI repository (9, 13). Wild-type optimized RSV F was sub-cloned into an expression vector and mutagenized by cysteine substitu-tions using the QuikChange II kit (Agilent, CA) and DNA oligonucleo-tides (IDT, TX). The cytoplasmic tail (CT) and the transmembranedomain (TM) of RSV F were swapped with hMPV F analog domains usingrecombinant DNA methods: the hRSV F sequence from amino acids 1 to524 (GenBank accession no. ACO83302.1) was joined to the hMPV Fsequence from amino acids 489 to 539 (GenBank accession no.AEK26895.1). Constructs were fully sequenced for quality verification.

Vaccine. RSV virus-like particles (VLPs) were produced in a suspen-sion culture of mammalian cells following transient transfection of theVLP protein expression plasmids. Expi293F cells were maintained in se-rum-free Expi293 expression medium (Life Technologies, CA). Thetransfection reaction mixture was assembled in separate tubes: 1.0 �g ofDNA vectors, 3.0 �g of 25-kDa linear polyethylenimine (PEI) (Poly-sciences, PA), and 0.1 ml of Opti-MEM, for each 1 ml of cell culture (2.5 �106 cells/ml). The plasmids for RSV F and hMPV M were transfected atratios of 70% and 30%, respectively. The mixture was incubated for 15min at room temperature and then added to the cells. Forty-eight hoursafter transfection, the culture supernatant was clarified by centrifugationat 8,000 � g for 10 min, and VLPs were concentrated by ultracentrifuga-tion for 2 h at 140,000 � g using an SW 28 rotor (Beckman Coulter, CA).The pellets were resuspended in 250 �l of buffer E (250 mM sucrose, 100mM MgSO4, 10 mM Tris [pH 7.5]) containing protease inhibitors(Thermo Scientific, MA). The VLPs were subsequently purified by ultra-centrifugation through a continuous sucrose gradient (10% to 50%) for 4h at 180,000 � g using an SW 40Ti rotor. The VLPs were collected fromthe 15% to 25% fraction and further purified by ultrafiltration using aVivaspin system with a 100-kDa cutoff (Sartorious, NY). The VLPs werecharacterized by Western blotting, dot blotting, and electron microscopy.The combo VLP vaccine was prepared by blending equal amounts (2 �g Fcontent) of VLPs with prefusion F and postfusion F per dose (4 �g total Fcontent). The formalin-inactivated RSV (FI-RSV) vaccine was producedusing the protocol described by Prince et al (14). The final preparation wasadjuvanted with Imject alum (Thermo Scientific, MA) to a final concen-tration of 4 mg/ml of aluminum hydroxide and then diluted 1:25 in phos-phate-buffered saline (PBS) for animal administration.

Dot blotting and Western blotting. The immune reactivity and anti-gen concentration of the VLP vaccine were assayed using a dot blot. TheVLP samples in a volume of 1 to 5 �l were absorbed into nitrocellulose for15 min and then blocked for 1 h with 3% nonfat dry milk in Tris-bufferedsaline plus 0.1% Tween 20 (TBS-Tween). The primary antibodies werepalivizumab (Synagis; MedImmune, MD), 131-2a (EMD-Millipore,MA), 5C4 (9), and anti-hMVP M (clone 4821; GeneTex, CA). Horserad-ish peroxidase (HRP)-conjugated mouse anti-human antibody or goatanti-mouse antibody was used as the secondary antibody. Detection was

FIG 1 Development of RSV F constructs using structural vaccinology. (A)Schematic representation of the wild-type (WT) RSV F primary structure. Fprotein matures by furin enzyme cleavage at sites I and II, generating the F2-F1protomer and releasing p27 glycopeptide. F protein is characterized by theheptad repeat domains HRA, HRB, and HRC, fusion peptide (FP), transmem-brane domain (TM), and cytosolic tail (CT), which is important for virionassembly with the matrix M protein. F elicits neutralizing antibodies able torecognize the antigenic sites: �, I, II, and IV. The figure includes a schematicpicture of the postfusion hybrid construct (Post) with the CT swapped with theanalogous domain of the hMPV F (green) and a schematic representation of

prefusion hybrid construct (Pre) with modification of disulfide bonds 102 to148 and 155 to 290. (B) Tridimensional structure representation of the Fprotomer in postfusion and prefusion conformations. The cysteine modifica-tions A102C and I148C are indicated in red. The prefusion conformation ismaintained by formation of a cysteine link between the F2 (purple) and F1(blue) chains (the positions of amino acids 102 and 148 are approximated inthe postfusion structure representation). (C) Dot blot analysis showing theimmune reactivity of recombinant RSV F proteins. The graph represents theresults from 3 independent experiments for 5C4 normalized over palivizumabimmune reactivity, and a single dot blot experiment is shown. The asteriskindicates a statistically significant P value (P � 0.05) between the postusionand prefusion conditions. ND, not detected.

Cimica et al.

452 cvi.asm.org June 2016 Volume 23 Number 6Clinical and Vaccine Immunology

on June 23, 2020 by guesthttp://cvi.asm

.org/D

ownloaded from

performed using enhanced chemiluminescence (ECL) substrate and adigital imaging system (FluorChem M system; ProteinSimple, CA). Thedot blot with palivizumab was used for quantifying the F content in thevaccines compared to a serial dilution of recombinant RSV F proteinstandard (Sino Biological, PA) using AlphaView SA imaging software(ProteinSimple). Western blotting was performed using a Novex gel sys-tem (Life Technologies) and NuPage 4 to 12% bis-Tris gel (Life Technol-ogies).

Electron microscopy analysis of the VLPs. The VLPs were examinedby negative staining and electron microscopy (Zeiss 902 transmissionelectron microscope with accelerating voltage of 80 kV). Five microlitersof resuspended particles was applied to a Formvar-coated copper grid(FCF300-Cu; Electron Microscopy Sciences, PA) and stained with 2.0%phosphotungstic acid (pH 7.0). Immunogold labeling was performed asfollows. First, 5 �l of resuspended samples was applied to a Formvar-coated grid and incubated for 5 min at room temperature (RT). Then thegrid was washed 5 times with buffer (0.1% fetal bovine serum [FBS] inPBS, 10 mM glycine, 0.01% NaN3), fixed with 4% paraformaldehyde inPBS for 15 min, and blocked in 1% bovine serum albumin (BSA) in PBSfor 30 min. The sample was floated onto 40 �l of palivizumab at a 1:1,000dilution in 0.1% BSA in PBS plus 0.01% NaN3 buffer overnight at 4°C.The grid was then placed in a humid chamber to prevent evaporation.Following 5 washes with buffer, the grid was placed on a drop of 6-nmgold beads conjugated with goat anti-human polyclonal IgG antibody(Abcam, MA) at a 1:3 dilution in 0.1% BSA in PBS plus 0.01% NaN3 for 2h at RT. Then the sample was washed 5 times in washing buffer andstained with 2.0% phosphotungstic acid (pH 7.0).

Virus production and titration. This study was performed in compli-ance with biosafety level 2 (BSL2) regulations. RSV strain A2 (ATCCVR-1540; ATCC, VA) was used for the murine challenge, production ofthe FI-RSV vaccine, immunization, and viral assays. RSV was propagatedin HEp-2 cells (ATCC CCL-23; ATCC) using Dulbecco’s modified Eagle’smedium (DMEM) with 2% FBS (Life Technologies) as the culture me-dium, according to the supplier’s protocol. Viral titration was performedby plaque assay using an HEp-2 monolayer. HEp-2 cells were infected for1 hour for viral absorption with 10-fold serially diluted virus in serum-freeDMEM, from a dilution range of 104 to 108. Subsequently, an overlay of1% methylcellulose (C-4888; Sigma-Aldrich, MO) in DMEM supple-mented with 2% FBS was applied to each well to prevent viral particlediffusion. After 4 to 5 days of incubation, the overlay was removed, andthe cells were fixed using cold methanol for 20 min at �20°C. Virusplaques were stained by immunocytochemical techniques using an anti-RSV goat antibody (AB1128; EMD-Millipore) diluted 1:500 in blockingbuffer as a primary antibody. Immune detection was performed using anHRP-conjugated rabbit anti-goat antibody (ab97105; Abcam, MA) as asecondary antibody. Immunostaining was performed using a DAB (3,3-diaminobenzidine) peroxidase (HRP) substrate kit (SK-4100; VectorLaboratories, CA), and plaques were counted using a light microscopewith a 4� to 20� lens objective.

Murine model for immunization and RSV infection. BALB/c mice(Mus musculus; 6-week-old females from Charles River) were housed atthe Department of Comparative Medicine, New York Medical College(Valhalla, NY). Mice were anesthetized with ketamine (100 mg/kg)-xyla-zine (10 mg/kg) administered via intraperitoneal injection before immu-nization or blood collection. Mice were immunized by intramuscular(i.m.) injections of 50 �l of the postfusion, prefusion, or combo VLPvaccine, the FI-RSV, and a placebo control (n � 10 per group). Immuni-zations were administered on days 1 and 14, and each VLP vaccine dosecontained 4 �g of total recombinant RSV F admixed in a 1:1 volume witha squalene-based oil-in-water nanoemulsion (AddaVax; InvivoGen, CA).The placebo group received PBS admixed with AddaVax at a 1:1 volume.Serum was collected by retro-orbital bleeding before and after immuni-zation. Mice were challenged with 1 � 106 PFU of RSV strain A2 con-tained in 50 �l administered via the intranasal route as small drops (Pi-petman with an ultraslim tip) at day 28. Groups of mice were euthanized

at 4 and 7 days postchallenge for lung and blood harvest (see Fig. S2 in thesupplemental material).

Pulmonary RSV quantification by plaque assay. Lungs from RSV-infected mice were harvested, weighed, and homogenized using an Omnitissue homogenizer (Omni International) in Opti-MEM I medium con-taining 25% sucrose, penicillin-streptomycin-glutamine (Life Technolo-gies), and 2.5 �g/ml amphotericin B (Fungizone) (Chem-Impex Interna-tional Inc., IL). Lung supernatants were obtained by centrifugation, andviral titers were measured by plaque assay as described above.

Neutralization assay. Plaque reduction neutralization tests (PRNT)were performed in duplicate using serum samples collected before viralchallenge (day 28). Serial dilutions of serum were incubated with 100 PFUof RSV strain A2 virus for 1 h at 37°C, and the neutralization power wasmeasured by a plaque assay in HEp-2 cells as described above. The 50%inhibitory concentration (IC50) calculation was performed by probit anal-ysis (15).

Luminex cytokine and chemokine analyses. The magnetic bead-based sandwich immunoassays for cytokines using the MILLIPLEX MAPmultiplex mouse cytokine panel 1 (EMD-Millipore) were performed ac-cording to the manufacturer’s instructions. The lung samples (25 �l) wereanalyzed in duplicate wells using a Luminex MagPix (Luminex Corp.,TX). The cytokine concentrations were determined by LuminexxPONENT v4.2 and EMD-Millipore MILLIPLEX Analyst v5.1 using 5-plog analysis. The gamma interferon (IFN-�) analysis in lung fluids wasconfirmed using an enzyme-linked immunosorbent assay (ELISA) kit(eBioscience, CA).

ELISA analysis of IgG subtypes. Each well of the ELISA plates (Corn-ing Costar, NY) was coated with RSV strain A2 containing 100 ng of Fprotein content determined by dot blot analysis using purified recombi-nant F protein (Sino Biological, PA) as the standard and incubated at 4°Covernight (see above). Serum samples were serially diluted in blockingbuffer (5% milk in TBS-Tween), applied in triplicate to the ELISA platesand incubated for 2 h at room temperature. Following washes, detectionwas carried out using the following antibodies: IgG1 subtype (115-035-205; Jackson ImmunoResearch Laboratory, PA) and IgG2a subtype (115-035-206; Jackson ImmunoResearch Laboratory). The ELISA measure-ments were performed using the ECL method and a microplate reader(Synergy H1; BioTek, VT). The endpoint calculation for the ELISA wasperformed according to the method of Frey et al. (16).

Histopathology. Lungs were harvested on day 4 postinfection andfixed in 10% buffered formalin phosphate. The lung samples were pro-cessed at the Department of Pathology, New York Medical College (Val-halla, NY) and stained with hematoxylin and eosin (H&E) following astandard protocol (17). The examination and scoring of the lung histopa-thology were performed by blind evaluation of the H&E slides.

Ethics statement. The Institutional Animal Care and Use Committee(IACUC) of the New York Medical College approved the protocol for themouse study. The study was performed in compliance with the approvedIACUC protocol (no. 33-2-0513H) and the Institutional Biosafety Com-mittee policies and under strict accordance to the Office of LaboratoryAnimal Welfare (OLAW) guidelines and the Public Health Service (PHS)Policy on Humane Care and Use of Laboratory Animals (NIH).

Statistics. Data were statistically analyzed and graphed using Graph-Pad Prism (GraphPad Software, CA), and errors bars represent the calcu-lated standard errors. The statistical significance of the data was measuredby one-way analysis of variance (ANOVA) with Dunnett’s multiple com-parisons between experimental conditions and t tests. The ratio of IgG2ato IgG1 was analyzed using the Taylor expansion statistical approach forcalculating standard errors. Pictures and images were processed usingAdobe Photoshop (Adobe Systems, CA).

RESULTSDevelopment of recombinant RSV F exhibiting postfusion andprefusion conformations. The structural analysis of the postfu-sion F shows that the C terminus of F2 is in the opposite orienta-

Respiratory Syncytial Virus-Like Particle Vaccine

June 2016 Volume 23 Number 6 cvi.asm.org 453Clinical and Vaccine Immunology

on June 23, 2020 by guesthttp://cvi.asm

.org/D

ownloaded from

tion and distant from the N terminus of F1 (Fig. 1B), whereasthese domains are adjacent in the prefusion conformation (9, 13).We identified within this region several amino acids that are inclose proximity, separated by �10 Å. Based on this analysis, wegenerated 9 recombinant constructs with alternative disulfidebonds between these domains to stabilize F in its prefusion con-formation and 1 with the furin cleavage site mutated (see Table S1in the supplemental material). Prefusion F VLPs were produced inmammalian cells and analyzed by dot blots with the monoclonalantibodies (mAbs) palivizumab (antigenic site II) to measure Fprotein expression and 5C4 to assess the presence and stability ofthe antigenic site � (see Table S1 and Fig. S1 in the supplementalmaterial). We evaluated the postfusion state with the mAb 131-2a,which is specific for the antigenic site I (see Table S1). We foundthat the most stable prefusion F contained an intrachain disulfidebond inside the F1 subunit described by McLellan et al. (S155C/S290C) (18) plus an interchain disulfide bond between the F1 andF2 subunits with the cysteine substitutions A102C and I148C (Fig.1A and B). By dot blot analysis with 5C4, we found that the pre-fusion F recombinant is recognized 19.2 2.4-fold more than thepostfusion F (Fig. 1C). To assess whether the disulfide bridgeA102C/I148C enhanced the stability of F prefusion, we tested anintermediate mutant having only one cysteine change (A102C).Indeed, the mutant A102C construct demonstrated reactivity with5C4 equivalent to that of the wild-type F postfusion construct. Inaddition, the dot blot analysis demonstrated that the combinationof disulfide bonds, S155C/S290C with A102C/I148C, enhancesthe 5C4 reactivity with respect to the single disulfide bond con-

structs (see Fig. S1). VLPs assembled with this mutant (S155C/S290C plus A102C/I148C) were used in the vaccine studies. Inaddition, this F construct contains the cytoplasmic tail domain ofhuman metapneumovirus (hMPV) F (Fig. 1A), which seems tofurther stabilize the structure of F incorporated in the particles asreflected by strong reactivity with 5C4 and palivizumab (Fig. 1C).This and other mutants without the hMPV F tail demonstratedstrong reactivity with 5C4 but weaker reactivity with palivizumab(see Table S1).

Generation of VLPs displaying RSV F postfusion or prefu-sion conformations. The RSV envelope displays three virally en-coded and membrane-anchored proteins, F, G, and SH (Fig. 2A)(19). Underlying the envelope is the matrix protein (M), whichduring morphogenesis multimerizes and drives virion assem-bly and budding (20). To assemble VLPs, we utilized the RSV Feither postfusion or prefusion together with the M of thehMPV as the scaffold, which we found to be more efficient thanthe RSV M in VLP formation (data not shown). To optimizethe interaction of RSV F and hMPV M, we replaced the cyto-plasmic domain of the RSV F protein with the analogous do-main of the hMPV F protein, which, based on yield analysis incomparison to the unmodified F, demonstrated enhanced RSVF recruitment and incorporation onto the VLP surface (Fig. 1Cand data not shown).

The analysis of purified VLPs by Western blotting showed thatthe RSV F hybrid copurified with the hMPV M (Fig. 2B) and thatreplacement of the cytoplasmic tail enhanced incorporation of theRSV F into particles compared to that for wild-type RSV F (data

FIG 2 Structure and morphology of RSV VLPs using recombinant RSV F and hMPV M. (A) Drawing of RSV viral particles. F, G, SH, M, M2, P, N, and L proteinsand genomic negative-sense RNA are indicated. (B) Western blot analysis of RSV VLPs using different antibodies such as goat anti-RSV (Anti-RSV Gt),palivizumab, and anti-hMPV M. The Western blotting experiment includes a purified RSV control. (C) Purified RSV VLPs were negatively stained and examinedby electron microscopy. The micrograph shows spherical and irregular particles (90 to 100 nm) decorated with surface projections or “spikes” (arrowheads)resembling the morphology of RSV F. (D) Electron micrograph of immunogold-labeled RSV VLPs probed with the humanized monoclonal antibody palivi-zumab and developed with a goat anti-human antibody coupled to gold spheres (10 nm). The detection of gold spheres demonstrates that F is decorating thesurface of the VLPs.

Cimica et al.

454 cvi.asm.org June 2016 Volume 23 Number 6Clinical and Vaccine Immunology

on June 23, 2020 by guesthttp://cvi.asm

.org/D

ownloaded from

not shown). These results suggest that the RSV F hybrids interactwith the hMPV M via its engineered cytoplasmic domain. Exam-ination of purified VLPs by electron microscopy (EM) showedspherical structures of 80 nm in diameter that display F spikesprotruding from the membrane envelope (Fig. 2C and D). Immu-nogold labeling for EM confirmed that the spikes were indeedcomposed of the glycoprotein F (Fig. 2D).

Efficacy evaluation of RSV F VLP vaccines in a murine model.To assess the protective efficacy of the VLP vaccines, BALB/c micewere immunized twice with formulations containing the post-fusion F VLP vaccine, the prefusion F VLP vaccine, or a combina-tion of the two VLP vaccines (combo VLP vaccine) and then chal-lenged with RSV (see Fig. S2 in the supplemental material). Toevaluate the safety of the VLP vaccines, we included a group ofmice immunized with the FI-RSV vaccine, expected to inducevaccine-enhanced disease. The analysis of the protective efficacyon day 4 postchallenge showed that the VLP-immunized micewere completely protected from RSV replication and did not showa detectable viral load (�50 PFU/gram of lung tissue), whereas theplacebo group demonstrated high levels of infective particles in-side their lungs (75,000 PFU/gram of lung tissue) (Fig. 3). Theassessment of the viral load on day 7 postchallenge demonstratedthe absence of replicating virus in the lungs of all of the animals(data not shown). To appraise the quality and magnitude of theantibody response, we measured the serum-neutralizing activityof the immunized animals prior to viral challenge (day 28) byplaque reduction neutralization tests (PRNT) (Fig. 3B). This anal-ysis showed that the combo VLP vaccine (postfusion plus prefu-sion F) elicited the highest level of neutralizing antibodies com-pared to either the postfusion or prefusion F single vaccineformulations. The prefusion F VLP vaccine, however, elicitedhigher neutralizing antibody titers than the postfusion vaccine,results that agree with previous reports (21, 22). On the otherhand, the FI-RSV vaccine failed to induce an appreciable level ofneutralizing antibodies. The palivizumab control demonstratedan IC50 neutralizing activity of 2 �g/ml, a value that is similar tothat for previous determinations (23).

VLP vaccine stimulates a balanced IgG response. We assessedthe magnitude of the IgG2a and IgG1 serum responses, which arecorrelates of Th1 and Th2 development, respectively (Fig. 3C andD). Mice that received the VLP vaccines demonstrated inductionof robust serum IgG responses compared to those receiving thepreimmune samples. On the other hand, serum from FI-RSV-immunized mice showed the highest level of IgG induction, sug-gesting that antibodies to multiple viral proteins were producedand detected in the whole-virus ELISA. The analysis of the IgGsubtype demonstrated that the combo VLP vaccine elicited a bal-anced Th1-mediated versus Th2-mediated response, associatedwith a greater ratio of IgG2a to IgG1 (Fig. 3D). As expected, theplacebo control demonstrated background levels of total and spe-cific IgGs against RSV.

Analysis of the cytokine profile in VLP-vaccinated and con-trol mice after RSV infection. We applied Luminex technology tostudy the cytokine and chemokine levels in lung homogenates ofVLP-vaccinated and control mice 4 days after challenge. We eval-uated the cytokine markers that correlate with the Th1, Th2, andTh17 types of immune responses as well as interleukin-10 (IL-10)(Fig. 4A to D). VLP-immunized mice showed a robust IFN-�response in comparison to that of the placebo control (Fig. 4A).On the other hand, immunization with the FI-RSV vaccine stim-

ulated a strong cytokine response that qualitatively and quantita-tively differed from the one elicited in mice immunized with theVLP vaccine or placebo. We found that FI-RSV immunizationinduced high levels of the cytokines IFN-�, tumor necrosis factoralpha (TNF-�), IL-4, IL-10, IL-17, and IL-1�, all of which havebeen associated with the exacerbation of RSV disease (24–29) (Fig.4A to D). In contrast, multiplex analysis of the placebo controldemonstrated very low expression levels of these cytokines, indi-cating that RSV replication did not trigger or perhaps curtailedproduction of these immune-signaling molecules (Fig. 4A to D).This is in agreement with the results of Lambert and coworkers(30), who also found very low levels of IFN-�, IL-5, IL-13, andIL-17 in the lungs of the placebo group on day 4 postchallenge.However, higher doses of the challenging virus (e.g., 1 � 107 PFU)increase the levels of the secreted cytokines as described by Ru-tigliano and Graham (29).

Chemokines recruit inflammatory cells to the infected tissueand are particularly elevated in bronchiolitis (31, 32). Thus, we

FIG 3 VLP vaccine protects against RSV infection. (A) Plaque assay analysisof viral titers in mouse lungs 4 days postchallenge shows undetectable viralreplication in VLP-vaccinated mice, whereas the placebo control groupdemonstrates a very productive infection. The dotted line indicates thelower detection limit. (B) The viral microneutralization assay shows the levelsof serum-neutralizing antibodies after vaccination. The combo VLP vaccina-tion resulted in a statistically significant enhancement of neutralizing antibod-ies with respect to VLP postfusion and prefusion vaccination (*, P � 0.05).ND, not detectable. (C) Measurement of RSV-specific IgG1 and IgG2a serumantibody titers. Serum samples collected on day 28 postimmunization (n � 4per group) were assayed for RSV-specific IgG isotypes by ELISA. The resultsare presented as the reciprocal of the log2 serum endpoint dilution. The anti-body titers in the vaccinated mice are statistically significant with respect to theplacebo control (*, P � 0.05). (D) Analysis of the ratio of IgG2a to IgG1demonstrates that the combo VLP vaccine induces a superior Th1-mediatedresponse. The results were generated using 4 mice for each condition.

Respiratory Syncytial Virus-Like Particle Vaccine

June 2016 Volume 23 Number 6 cvi.asm.org 455Clinical and Vaccine Immunology

on June 23, 2020 by guesthttp://cvi.asm

.org/D

ownloaded from

analyzed eotaxin, monocyte chemoattractant protein 1 (MCP-1),macrophage inflammatory protein 1 alpha (MIP-1�), andRANTES, all of which are involved in lung immune cell infiltra-tion during bronchiolitis (Fig. 5). While FI-RSV vaccinestrongly augmented each chemokine, the VLP-immunized an-imals had a less significant induction of these inflammatorymediators, the levels of which were closer to those seen in theplacebo controls.

Evaluation of VLP vaccine safety by lung histopathology ex-amination. VLP vaccine safety and tolerability were further eval-uated by histological examination of the lung tissues of VLP-vac-cinated mice at 4 days postinfection (Fig. 6A and B). The placebocontrol mice that received a primary infection with 106 PFU ofRSV experienced some interstitial cellular infiltrate but limited orno sign of perivascular infiltration on day 4 postchallenge, indi-cating that virus replication was tolerated without provoking se-

rious lesions. Indeed, a previous report (30) has shown minimaleosinophilic infiltration in primary infected mice under the sameexperimental conditions.

FI-RSV-vaccinated mice displayed massive perivascular, peri-bronchial, and interstitial infiltrations of inflammatory cells. Incontrast, the VLP-immunized mice showed limited immune cellinfiltrations. Blinded scoring of the perivascular infiltration dem-onstrated that the combo VLP vaccine had the lowest level ofhistological changes (Fig. 6B). The perivascular infiltrations in thepostfusion and prefusion vaccine groups were more pronouncedthan those of the placebo controls; however, the overall lung ar-chitectures were not significantly different among these groups.Severe RSV disease and FI-RSV vaccine-enhanced disease arecharacterized by cellular infiltration and lung hyperinflation (33,34). We found that the lungs from mice immunized with FI-RSV were significantly larger and were 30% heavier than the

FIG 4 Cytokine responses in VLP-vaccinated mouse lungs after RSV infection. (A) The cytokines for the Th1-mediated response are IFN-�, IL-12p40, andTNF-�. (B) The Th2-mediated response cytokine measurement includes IL-4, IL-5, and IL-13. (C) The IL-17 and IL-1� analysis indicates the Th17-mediatedresponse. (D) The IL-10 cytokine analysis demonstrates immunoregulatory process development. The results were generated using a group of 4 mice for eachcondition at 4 days postinfection. An asterisk indicates a statistically significant difference between the experimental condition and the placebo control;differences between the FI-RSV and VLP vaccines were statistically significant (P � 0.05) in all groups, with the exception of the prefusion group in the IFN-�assay. ND, not detectable.

Cimica et al.

456 cvi.asm.org June 2016 Volume 23 Number 6Clinical and Vaccine Immunology

on June 23, 2020 by guesthttp://cvi.asm

.org/D

ownloaded from

lungs from the placebo controls, while the lungs from the VLP-vaccinated mice did not differ from those of the placebo group(Fig. 6C).

DISCUSSION

Here we present data on the immunogenicity, efficacy, and safetyof a novel VLP-based RSV vaccine constructed with different con-formations of the RSV F glycoprotein. Previously tested subunitvaccines were formulated primarily with RSV F in its postfusionconformation (35, 36). Recent studies (21, 22, 37), however, havedemonstrated that the RSV F in the prefusion conformation hasthe ability to elicit higher levels of neutralizing antibodies than thepostfusion conformation. Indeed, different groups have shownthat vaccines containing RSV F in the prefusion conformationwere superior for protecting mice and cotton rats from RSV infec-tion (21, 22, 37). Considering these data, we designed, produced,and tested a novel recombinant stabilized prefusion F that ishighly expressed in mammalian cells, is incorporated into VLPs,and is recognized strongly by the mAb 5C4, which binds the pre-fusion-only antigenic site �.

Although VLPs are strong immunogens, we included an adju-vant in the VLP vaccine in order to elicit the greatest immunoge-nicity. Protection against RSV infection may require a greater im-munity than that stimulated by natural infection, which does notprevent reinfection. We selected the squalene-based oil emulsionbecause it is a potent inducer of both Th1- and Th2-mediatedimmunity and is well tolerated and safe (38, 39).

Assessment of the protective efficacy afforded by VLP vaccina-tion after RSV challenge showed that each one of the three VLPvaccine formulations protected the lungs from viral infection.Furthermore, evaluation of the serum neutralization potencyshowed that the prefusion F VLP vaccine induced antibodies with

FIG 5 Chemokine responses in VLP-vaccinated mouse lungs after RSV infec-tion. Analyses include eotaxin, MCP-1, MIP-1�, and RANTES. The resultswere generated using a group of 4 mice for each condition at 4 days postinfec-tion. An asterisk indicates a statistically significant difference between an ex-perimental condition and the placebo control (P � 0.05). Differences betweenthe FI-RSV and the VLP vaccines are also statistically significant (P � 0.05) inall groups.

FIG 6 RSV VLP vaccines do not induce “vaccine-enhanced disease” in murine lung. (A) Hematoxylin and eosin staining of mouse lung 4 days after viralchallenge. Microscopic examination of the lungs after the combo VLP vaccination shows the absence of or minor pulmonary pathology. On the other hand, theFI-RSV vaccination induces a very high perivascular immune infiltration. A placebo control is included as a reference. (B) Perivascular infiltration was scored byblind evaluation of sections of mouse lung stained with hematoxylin and eosin 4 days after the viral challenge. The scores range from 0 for normal to a maximumof 3 for massive infiltration. An asterisk indicates a statistically significant difference between VLP vaccination and FI-RSV (P � 0.05). (C) Mouse lungs wereharvested 4 and 7 days after RSV challenge and weighed. An asterisk indicates a statistically significant difference with respect to the VLP vaccines and placebocontrol (P � 0.05).

Respiratory Syncytial Virus-Like Particle Vaccine

June 2016 Volume 23 Number 6 cvi.asm.org 457Clinical and Vaccine Immunology

on June 23, 2020 by guesthttp://cvi.asm

.org/D

ownloaded from

a higher neutralization power than did immunization with thepostfusion F VLP vaccine, in agreement with the results of previ-ous studies (21, 22), although not to the same extent. However,the combo VLP vaccine showed the best neutralization activity,reaching a neutralizing power that was 4-fold greater than thatseen with the postfusion VLP vaccine and 2-fold greater thanthat seen with the prefusion VLP vaccine. Notably, all VLP vac-cines contained the same total F protein content (4 �g total),suggesting that the combination of the two conformations of Fprotein may be synergistic in eliciting a protective immune re-sponse. Recent studies performed with the Newcastle disease virusVLP vaccines showed similar results when the postfusion and pre-fusion forms were compared; however, a combo formulation wasnot tested (21, 40). It seems reasonable to speculate that thecombo VLP vaccine displays a larger repertoire of neutralizingepitopes than either of its components. This outcome is significantfor RSV vaccine development and requires further investigation todefine the underlying mechanism.

Consistent with the antibody neutralization assay data, the IgGisotyping analysis showed that the combo VLP vaccine induced abalanced Th1-mediated immune response. On the other hand, theFI-RSV vaccine elicited high levels of IgG that did not correlatewith the induction of neutralizing antibodies. This outcomeclearly illustrates the dichotomy between high antibody titers andneutralizing capacity, which has been the hallmark of the FI-RSVvaccine-enhanced disease.

The cytokine analyses showed that all VLP-vaccinated miceproduced statistically significant levels of IFN-� compared tothose of the placebo group, which is a correlate of induction of aTh1 type of immune response. On the other hand, the FI-RSVvaccine induced high levels of IFN-� and TNF-�, as well as highproduction of Th2-polarizing cytokines (IL-4, IL-5, and IL-13),Th17-polarizing cytokines (IL-17 and IL-1�) and IL-10, andchemokines (eotaxin, MCP-1, MIP-1�, and RANTES), all ofwhich have been associated with enhanced RSV pathogenesis (24–29). In contrast, VLP-vaccinated mice produced much smalleramounts of these inflammatory cytokines and chemokines. Fur-thermore, histopathology studies showed that VLP vaccinationdid not induce the detrimental immune cell infiltration inside thelung and that the combo VLP formulation was, in this regard, thebest-tolerated vaccine.

In summary, we describe the production of RSV VLPs com-posed of RSV F glycoprotein that display different epitopes suit-able for the elicitation of neutralizing antibodies. Considering thediversity, neutralizing strength, and distribution of epitopes, bothshared and unique, between the two conformations of F, it seemedimportant to compare the immunogenicity and efficacy of singleVLP vaccines (postfusion or prefusion F) with those of a comboVLP formulation. This study showed that the combo VLP vac-cine, composed of the multiple epitopes revealed in the post-fusion and prefusion F, afforded complete protection againstRSV infection and elicited production of the highest level ofserum-neutralizing antibodies that correlate with the develop-ment of a strong Th1 immune response. Furthermore, immu-nization with this vaccine proved to be safe, a condition thatmust be satisfied by any RSV vaccine candidate. We anticipatethat the combo VLP vaccine may elicit a broader spectrum ofneutralizing antibodies and thus afford better protectionagainst RSV infection and is a viable safe and efficacious can-didate for clinical evaluation.

ACKNOWLEDGMENTS

We thank Barney S. Graham from NIH for the generous donation of the5C4 antibody. We thank George Martin and Katherine Frederick for help-ful discussions and critical review of the manuscript.

This work was supported by a Small Business Innovation Research(SBIR) grant from the National Institutes of Health (NIH-R43AI108099)to J.M.G. The electron microscopy studies were performed in the labora-tory of Paul Gottlieb supported by the National Institute of General Med-ical Science under grant SC1-GM092781 and the National Institutes ofHealth Research Centers in Minority Institutions (NIH/NCRR/RCMI)CCNY/grant 5G12MD007603-30.

V.C., H.B., and J.M.G. are employees of TechnoVax, Inc., and B.B. isan employee of EMD-Millipore Corporation. V.C., H.B., and J.M.G. arethe inventors on a patent related to this publication.

FUNDING INFORMATIONThis work, including the efforts of Velasco Cimica, Héléne Boigard, andJose M. Galarza, was funded by HHS | National Institutes of Health (NIH)(NIH-R43AI108099). This work, including the efforts of Alexandra Ali-mova and Paul Gottlieb, was funded by HHS | National Institutes ofHealth (NIH) (5G12MD007603-30). This work, including the efforts ofAlexandra Alimova and Paul Gottlieb, was funded by HHS | NIH | Na-tional Institute of General Medical Sciences (NIGMS) (SC1-GM092781).

REFERENCES1. Meng J, Stobart CC, Hotard AL, Moore ML. 2014. An overview of

respiratory syncytial virus. PLoS Pathog 10:e1004016. http://dx.doi.org/10.1371/journal.ppat.1004016.

2. Hall CB, Weinberg GA, Iwane MK, Blumkin AK, Edwards KM, StaatMA, Auinger P, Griffin MR, Poehling KA, Erdman D, Grijalva CG, ZhuY, Szilagyi P. 2009. The burden of respiratory syncytial virus infection inyoung children. N Engl J Med 360:588 –598. http://dx.doi.org/10.1056/NEJMoa0804877.

3. Knudson CJ, Varga SM. 2015. The relationship between respiratorysyncytial virus and asthma. Vet Pathol 52:97–106. http://dx.doi.org/10.1177/0300985814520639.

4. Wu P, Hartert TV. 2011. Evidence for a causal relationship betweenrespiratory syncytial virus infection and asthma. Expert Rev Anti InfectTher 9:731–745. http://dx.doi.org/10.1586/eri.11.92.

5. Johnson KM, Bloom HH, Mufson MA, Chanock RM. 1962. Naturalreinfection of adults by respiratory syncytial virus. Possible relation tomild upper respiratory disease. N Engl J Med 267:68 –72.

6. Falsey AR, Singh HK, Walsh EE. 2006. Serum antibody decay in adultsfollowing natural respiratory syncytial virus infection. J Med Virol 78:1493–1497. http://dx.doi.org/10.1002/jmv.20724.

7. Walsh EE, Falsey AR. 2012. Respiratory syncytial virus infection in adultpopulations. Infect Disord Drug Targets 12:98 –102. http://dx.doi.org/10.2174/187152612800100116.

8. Modjarrad K, Giersing B, Kaslow DC, Smith PG, Moorthy VS, WHORSV Vaccine Consultation Expert Group. 2016. WHO consultation onRespiratory Syncytial Virus Vaccine Development Report from a WorldHealth Organization meeting held on 23�24 March 2015. Vaccine 34:190�197. http://dx.doi.org/10.1016/j.vaccine.2015.05.093.

9. McLellan JS, Chen M, Leung S, Graepel KW, Du X, Yang Y, Zhou T,Baxa U, Yasuda E, Beaumont T, Kumar A, Modjarrad K, Zheng Z,Zhao M, Xia N, Kwong PD, Graham BS. 2013. Structure of RSV fusionglycoprotein trimer bound to a prefusion-specific neutralizing antibody.Science 340:1113–1117. http://dx.doi.org/10.1126/science.1234914.

10. Magro M, Mas V, Chappell K, Vázquez M, Cano O, Luque D, TerrónMC, Melero JA, Palomo C. 2012. Neutralizing antibodies against thepreactive form of respiratory syncytial virus fusion protein offer uniquepossibilities for clinical intervention. Proc Natl Acad Sci U S A 109:3089 –3094. http://dx.doi.org/10.1073/pnas.1115941109.

11. Corti D, Bianchi S, Vanzetta F, Minola A, Perez L, Agatic G, GuarinoB, Silacci C, Marcandalli J, Marsland BJ, Piralla A, Percivalle E, SallustoF, Baldanti F, Lanzavecchia A. 2013. Cross-neutralization of fourparamyxoviruses by a human monoclonal antibody. Nature 501:439 –443. http://dx.doi.org/10.1038/nature12442.

12. Gilman MSA, Moin SM, Mas V, Chen M, Patel NK, Kramer K, Zhu Q,Kabeche SC, Kumar A, Palomo C, Beaumont T, Baxa U, Ulbrandt ND,

Cimica et al.

458 cvi.asm.org June 2016 Volume 23 Number 6Clinical and Vaccine Immunology

on June 23, 2020 by guesthttp://cvi.asm

.org/D

ownloaded from

Melero JA, Graham BS, McLellan JS. 2015. Characterization of a prefu-sion-specific antibody that recognizes a quaternary, cleavage-dependentepitope on the RSV fusion glycoprotein. PLoS Pathog 11:e1005035. http://dx.doi.org/10.1371/journal.ppat.1005035.

13. McLellan JS, Yang Y, Graham BS, Kwong PD. 2011. Structure of respi-ratory syncytial virus fusion glycoprotein in the postfusion conformationreveals preservation of neutralizing epitopes. J Virol 85:7788 –7796. http://dx.doi.org/10.1128/JVI.00555-11.

14. Prince GA, Curtis SJ, Yim KC, Porter DD. 2001. Vaccine-enhancedrespiratory syncytial virus disease in cotton rats following immunizationwith Lot 100 or a newly prepared reference vaccine. J Gen Virol 82:2881–2888. http://dx.doi.org/10.1099/0022-1317-82-12-2881.

15. Bliss CI. 1934. The method of probits. Science 79:38 –39. http://dx.doi.org/10.1126/science.79.2037.38.

16. Frey A, Di Canzio J, Zurakowski D. 1998. A statistically defined endpointtiter determination method for immunoassays. J Immunol Methods 221:35– 41. http://dx.doi.org/10.1016/S0022-1759(98)00170-7.

17. Moreno PR, Palacios IF, Leon MN, Rhodes J, Fuster V, Fallon JT. 1999.Histopathologic comparison of human coronary in-stent and post-balloon angioplasty restenotic tissue. Am J Cardiol 84:462– 466. http://dx.doi.org/10.1016/S0002-9149(99)00334-3.

18. McLellan JS, Chen M, Joyce MG, Sastry M, Stewart-Jones GBE, Yang Y,Zhang B, Chen L, Srivatsan S, Zheng A, Zhou T, Graepel KW, KumarA, Moin S, Boyington JC, Chuang G-Y, Soto C, Baxa U, Bakker AQ,Spits H, Beaumont T, Zheng Z, Xia N, Ko S-Y, Todd J-P, Rao S,Graham BS, Kwong PD. 2013. Structure-based design of a fusion glyco-protein vaccine for respiratory syncytial virus. Science 342:592–598. http://dx.doi.org/10.1126/science.1243283.

19. Liljeroos L, Krzyzaniak MA, Helenius A, Butcher SJ. 2013. Architectureof respiratory syncytial virus revealed by electron cryotomography. ProcNatl Acad Sci U S A 110:11133–11138. http://dx.doi.org/10.1073/pnas.1309070110.

20. Ghildyal R, Ho A, Jans DA. 2006. Central role of the respiratory syncytialvirus matrix protein in infection. FEMS Microbiol Rev 30:692–705. http://dx.doi.org/10.1111/j.1574-6976.2006.00025.x.

21. McGinnes Cullen L, Schmidt MR, Kenward SA, Woodland RT, Mor-rison TG. 2015. Murine immune responses to virus-like particle-associated pre- and postfusion forms of the respiratory syncytial virus Fprotein. J Virol 89:6835– 6847. http://dx.doi.org/10.1128/JVI.00384-15.

22. Liang B, Surman S, Amaro-Carambot E, Kabatova B, Mackow N,Lingemann M, Yang L, McLellan JS, Graham BS, Kwong PD, Schaap-Nutt A, Collins PL, Munir S. 2015. Enhanced neutralizing antibodyresponse induced by respiratory syncytial virus pre-fusion F protein ex-pressed by a vaccine candidate. J Virol 89:9499�9510. http://dx.doi.org/10.1128/JVI.01373-15.

23. Schickli JH, Whitacre DC, Tang RS, Kaur J, Lawlor H, Peters CJ,Jones JE, Peterson DL, McCarthy MP, Van Nest G, Milich DR. 2015.Palivizumab epitope-displaying virus-like particles protect rodentsfrom RSV challenge. J Clin Invest 125:1637–1647. http://dx.doi.org/10.1172/JCI78450.

24. Arruvito L, Raiden S, Geffner J. 2015. Host response to respiratorysyncytial virus infection. Curr Opin Infect Dis 28:259 –266. http://dx.doi.org/10.1097/QCO.0000000000000159.

25. Christiaansen AF, Knudson CJ, Weiss KA, Varga SM. 2014. The CD4 Tcell response to respiratory syncytial virus infection. Immunol Res 59:109 –117. http://dx.doi.org/10.1007/s12026-014-8540-1.

26. Stoppelenburg AJ, de Roock S, Hennus MP, Bont L, Boes M. 2014.Elevated Th17 response in infants undergoing respiratory viral infection.Am J Pathol 184:1274 –1279. http://dx.doi.org/10.1016/j.ajpath.2014.01.033.

27. Durbin JE, Johnson TR, Durbin RK, Mertz SE, Morotti RA, Peebles RS,

Graham BS. 2002. The role of IFN in respiratory syncytial virus patho-genesis. J Immunol 168:2944 –2952. http://dx.doi.org/10.4049/jimmunol.168.6.2944.

28. Sheeran P, Jafri H, Carubelli C, Saavedra J, Johnson C, Krisher K,Sánchez PJ, Ramilo O. 1999. Elevated cytokine concentrations in thenasopharyngeal and tracheal secretions of children with respiratory syn-cytial virus disease. Pediatr Infect Dis J 18:115–122. http://dx.doi.org/10.1097/00006454-199902000-00007.

29. Rutigliano JA, Graham BS. 2004. Prolonged production of TNF-alphaexacerbates illness during respiratory syncytial virus infection. J Immunol173:3408 –3417. http://dx.doi.org/10.4049/jimmunol.173.5.3408.

30. Lambert SL, Aslam S, Stillman E, MacPhail M, Nelson C, Ro B,Sweetwood R, Lei YM, Woo JC, Tang RS. 2015. A novel respiratorysyncytial virus (RSV) F subunit vaccine adjuvanted with GLA-SE elicitsrobust protective TH1-type humoral and cellular immunity in rodentmodels. PLoS One 10:e0119509. http://dx.doi.org/10.1371/journal.pone.0119509.

31. Miller AL, Bowlin TL, Lukacs NW. 2004. Respiratory syncytial virus-induced chemokine production: linking viral replication to chemokineproduction in vitro and in vivo. J Infect Dis 189:1419 –1430. http://dx.doi.org/10.1086/382958.

32. McNamara PS, Flanagan BF, Hart CA, Smyth RL. 2005. Production ofchemokines in the lungs of infants with severe respiratory syncytial virusbronchiolitis. J Infect Dis 191:1225–1232. http://dx.doi.org/10.1086/428855.

33. Henry RL, Hodges IG, Milner AD, Stokes GM. 1983. Respiratory prob-lems 2 years after acute bronchiolitis in infancy. Arch Dis Child 58:713–716. http://dx.doi.org/10.1136/adc.58.9.713.

34. De Swart RL, Kuiken T, Timmerman HH, van Amerongen G, Van DenHoogen BG, Vos HW, Neijens HJ, Andeweg AC, Osterhaus ADME.2002. Immunization of macaques with formalin-inactivated respiratorysyncytial virus (RSV) induces interleukin-13-associated hypersensitivityto subsequent RSV infection. J Virol 76:11561–11569. http://dx.doi.org/10.1128/JVI.76.22.11561-11569.2002.

35. Swanson KA, Settembre EC, Shaw CA, Dey AK, Rappuoli R, MandlCW, Dormitzer PR, Carfi A. 2011. Structural basis for immunizationwith postfusion respiratory syncytial virus fusion F glycoprotein (RSV F)to elicit high neutralizing antibody titers. Proc Natl Acad Sci U S A 108:9619 –9624. http://dx.doi.org/10.1073/pnas.1106536108.

36. Raghunandan R, Lu H, Zhou B, Xabier MG, Massare MJ, Flyer DC,Fries LF, Smith GE, Glenn GM. 2014. An insect cell derived respiratorysyncytial virus (RSV) F nanoparticle vaccine induces antigenic site II an-tibodies and protects against RSV challenge in cotton rats by active andpassive immunization. Vaccine 32:6485– 6492. http://dx.doi.org/10.1016/j.vaccine.2014.09.030.

37. Krarup A, Truan D, Furmanova-Hollenstein P, Bogaert L, Bouchier P,Bisschop IJM, Widjojoatmodjo MN, Zahn R, Schuitemaker H, McLel-lan JS, Langedijk JPM. 2015. A highly stable prefusion RSV F vaccinederived from structural analysis of the fusion mechanism. Nat Commun6:8143. http://dx.doi.org/10.1038/ncomms9143.

38. Ott G, Barchfeld GL, Chernoff D, Radhakrishnan R, van Hoogevest P,Van Nest G. 1995. MF59. Design and evaluation of a safe and potentadjuvant for human vaccines. Pharm Biotechnol 6:277–296.

39. O’Hagan DT, Ott GS, Nest GV, Rappuoli R, Giudice GD. 2013. Thehistory of MF59 adjuvant: a phoenix that arose from the ashes. Expert RevVaccines 12:13–30. http://dx.doi.org/10.1586/erv.12.140.

40. Cullen LM, Blanco JCG, Morrison TG. 2015. Cotton rat immune re-sponses to virus-like particles containing the pre-fusion form of respira-tory syncytial virus fusion protein. J Transl Med 13:350. http://dx.doi.org/10.1186/s12967-015-0705-8.

Respiratory Syncytial Virus-Like Particle Vaccine

June 2016 Volume 23 Number 6 cvi.asm.org 459Clinical and Vaccine Immunology

on June 23, 2020 by guesthttp://cvi.asm

.org/D

ownloaded from